Originally Posted by chaos
Do have to admit, I like Jeri and her work
No one says something else - i like her - she is nice person, attractive girl, very positive due of being female which is uncommon for electronics - probably her popularity is related highly to this that she is a SHE not HE - but as expressed this majsta - everyone who at least trying to do something with FPGA realize that this is a bit trickier than hookup few tens wires to board. I think that perhaps dev board act as logic analyzer (SignalTap) - this is fine for me, perhaps as basic state machine with some Agnus capabilities (Agnus not present in socket), OK but defintelly this is not eneough to emulate each chip. Ask Jens about estimated FPGA requirements - Clone-A seem to be cycle exact replica for Amiga chips thus this should be very close to schematics and will provide quite correct figures.
MC68000 is different example - MC68k is approx 68000 transistors but why on FPGA this softcore use more than few hundred thousands transistor? FPGA have different architecture - some block are not fully utilized etc. There are even online available patents (
http://www.freepatentsonline.com/4325121.pdf http://www.freepatentsonline.com/4296469.pdf http://www.freepatentsonline.com/4307445.pdf
) with quasi detailed fuse maps for PLA used in MC68000 but i doubt that they help reduce size of used LE in FPGA.
And yes - i agree - Minimig is probably large due of reimplementation HW form SW point of view - probably far from being efficient but still it works more or less OK ans can be used as Amiga alternative for all users without real Amiga.
What i expect is FPGA cycle exact replica, with new features added to OCS/ECS/AGA chips, with SDRAM used where classic transfers are use only fraction of a time and CHIP is available fully in almost independent way, with some new features like different data organization, perhaps some multimedia acceleration etc etc etc